Researchu On Traffic Sign Recognition Algorithm Based On Amplitude Spectrum And Neural Network

Traffic Sign Recognition (TSR) system is an important part of the intelligent car, it collects and recognizes the traffic sign information appears during it’s driving in order to make direction or warning. So in the automotive vision systems, how to effectively identify road traffic signs is a very important research topic. Although many algorithms for recognition of traffic signs are developed and they play the important roles in some practice instances, they also have some essential drawbacks and deficiencies. So it is important to study the refinement algorithms with the more real-time and better precision using the new developed techniques and theories for their applications.In this paper, methods of localization and segmentation of the traffic sign and methods of traffic sign automatic recognition are presented.In the aspect of traffic sign localization, a new image localization algorithm is proposed, which is algorithm of secondary image localization based on HSI color space. Simulation results show that this new algorithm can achieve high localization accuracy in almost every location occasion.In the aspect of traffic sign automatic recognition, feature extraction based on amplitude spectrum and moment invariants and a double layer networks based on BP network are proposed. Simulation results show that the seven features are not only translation invariant and rotation invariant but also insensitive to noise and illumination. Those algorithms can achieve high localization accuracy and recognition rate in almost every complex background, and have important application prospects.